Back to News
Market Impact: 0.35

3 AI Stocks to Still Buy If Inflation Stays Sticky

+1
Artificial IntelligenceInflationInterest Rates & YieldsCompany FundamentalsAnalyst EstimatesInvestor Sentiment & Positioning

U.S. inflation rose to 4.2% y/y in May (highest in three years) and remains well above the Fed’s 2% target, raising the risk the Fed may need to lift rates again. Despite the potential headwind for high-growth AI equities, the article argues Nvidia, CoreWeave, and Broadcom remain attractive on dips, citing outsized growth expectations (e.g., Nvidia revenue/EPS CAGRs of 46% for FY2026–FY2029; Broadcom AI chip sales up 65% to $20B in FY2025 and expected to reach at least $100B by FY2027). Overall, it frames the market pullback risk as short-term while emphasizing strong AI fundamentals and valuation support.

Analysis

This is more a duration/positioning call than a fundamental reset: the higher-rate backdrop mainly punishes the parts of AI where revenue is farthest out and financing dependence is highest. That makes CRWV the cleanest loser on a sustained yield backup because its equity value is effectively a call option on continued capital-market access and customer concentration, while NVDA and AVGO are better insulated by already-monetizing end demand and pricing power.

The second-order winner is not just “AI hardware” broadly, but the vendors with the tightest control over deployment economics. If hyperscalers keep leaning into custom silicon, AVGO can take share in inference even if GPU demand stays strong; that creates a structural split between training-heavy exposure and inference-optimized exposure. AMD remains vulnerable to multiple compression because it needs execution to justify a higher share in a market where buyers increasingly favor integrated ecosystems over standalone chips.

Near term, the key catalyst is not inflation itself but the next move in real yields and the next hyperscaler capex commentary window. If rates back up further over the next 1-3 months, expect the market to punish unprofitable AI infra first and ask tougher questions about payback periods; over 6-18 months, the real test is whether AI workloads become cash-generative enough to sustain current spending without balance-sheet subsidy. A cooler inflation print or dovish Fed pivot would reverse this trade quickly, especially for CRWV and high-multiple semis.

The contrarian miss is that the market may already be over-penalizing quality AI compounders for macro duration risk while underpricing the scarcity of their supply. In other words, higher rates can compress multiples, but they do not easily create new GPU supply or new customer lock-in; that argues for buying the leaders on weakness rather than chasing beta across the entire AI basket.

More News